StatsInvQpCDF
StatsInvQpCDF (ng, nt, df, alpha, side, sSizeWave)
The StatsInvQpCDF function returns the critical value of the Q' cumulative distribution function for ng the number of groups, nt the number of treatments, and df the error degrees of freedom. side =1 for upper-tail or side =2 for two-tailed critical values.
sSizeWave is an integer wave of ng columns and nt rows specifying the number of samples in each treatment. If sSizeWave is a null wave ($""), StatsInvQpCDF computes the number of samples from:
df = ng * nt * (n - 1)
with n truncated to an integer.
Details
StatsInvQpCDF is a modified Q distribution typically used with Dunnett's test, which compares the various means with the mean of the control group or treatment.
StatsInvQpCDF differs from other StatsInvXXX functions in that you do not specify a cdf value for the inverse (usually 1-alpha for the critical value). Here alpha selects one- or two-tailed critical values.
It is computationally intensive, taking longer to execute for smaller alpha values.
Examples
The critical value for a Dunnett test comparing 4 treatments with 4 samples and (upper tail) 0.05 significance is:
// n=4 because 12=1*4*(4-1).
Print StatsInvQpCDF(1,4,12,0.05,1,$"")
2.28734